key: cord-1046431-n2iv0k3x
authors: Wang, Qiang; Wang, Shasha
title: Preventing carbon emission retaliatory rebound post-COVID-19 requires expanding free trade and improving energy efficiency
date: 2020-07-21
journal: Sci Total Environ
DOI: 10.1016/j.scitotenv.2020.141158
sha: 2a5290cdbad2b75e11cbe0282425d71ce4ac90f5
doc_id: 1046431
cord_uid: n2iv0k3x

Abstract Existing studies have shown that the COVID-19 pandemic caused a sharp drop in carbon emissions in 2020. A recent example of the impact of sudden extreme events on carbon emissions occurred in the 2008 global financial crisis, in which carbon emissions dipped in 2009, but jumped in 2010. This study is aimed to discuss how to preventing the retaliatory growth of carbon emissions post COVID-19 through learning the lessons from analysis of short-term and long-term drivers of carbon emissions. This study explored the short-term (annual) effects (population scale. Affluence level, carbon intensity, energy intensity) of changes in carbon emissions by decomposing carbon emissions in the world, different income groups and selected countries before and after the 2008 financial crisis using LMDI technique. In addition, this study explored the long-term effects (energy consumption per capita, energy structure, energy intensity, foreign direct investment, and trade openness) of changes in carbon emissions by decomposing carbon emission in the world and different income groups from 1990 to 2014 using VAR technique. The decomposition results of short-term drivers of carbon emission uncovered that the improvement in energy efficiency (decline in energy intensity) was the main reason for the retaliatory rebound in carbon emissions post-2008 financial crisis, especially in high-income countries. The decomposition results of long-term drivers of carbon emission uncovered that trade openness contributed to reduce carbon emission in the world and the all the incomes groups in the long term, although trade openness led to increase in carbon emission in developing countries in the short term. To prevent retaliatory rebound of carbon emissions, what we should learn two lessons from the decomposition of carbon emission: improving energy efficiency, and expanding trade openness. Unfortunately, energy efficiency has been neglected in the economic recovery plans to respond to COVID-19 of various countries, especially developed countries, and worse, trade protectionism is on the rise, especially in developed countries. Therefore, we are pessimistic about preventing a retaliatory rebound in carbon emissions post-COVID-19 for now.

The COVID-19 pandemic caused severe impact on public health and shocked global economy. In addition, the outbreak of COVID-19 had a serious impact on environment, more specifically on carbon emission. Le Quéré et al.(Le Quéré et al., 2020) found forced confinement has an important impact on global carbon emission changes. They estimated that the confinement is to decline global carbon emission by early April 2020 by 17% compared with 2019 mean level. As for carbon emission in the rest of the year, it depends on the duration and extent of the confinement. Jeff

Tollefson analyzed several studies on carbon emission changes during the COVID-19m and he found carbon emission is bound to decline more than one billion tons in the first months compared with the same period in 2019 (Tollefson, 202) . A report from Carbon Brief indicates that COVID-19 is likely to cause the largest annual fall in carbon emission as more countries enforcing lockdowns to stop this pandemic (Carbon Brief, 2020a) . Then another report from Carbon Brief predicts that carbon emission probably see a decline up to 2729 million tons carbon emission in 2020 as a whole, which is the first one to quantify carbon emission changes on a daily basis (Carbon Brief, 2020b) . Moreover, IEA(International Energy Agency, 2020) forecasted a decrease of 8% in 2020 carbon emission. Besides, Evans (Evans, 2020) believed that as work resumption and economy stimulation, energy consumption will surge and carbon emission will rebound sharply. EIA estimated that carbon emission will decrease by 12.2% in 2020 and increase by 6.0% in 2021 in the United J o u r n a l P r e -p r o o f 6 situation . The United Nations Framework Convention on Climate Change (UNFCCC) promoted a "quantified emissions limitation and reduction objective" (Rogelj et al., 2019; Schleussner et al., 2016) . However, in late 2019, a new pandemic emerged and rapidly spread out, which will definitely threaten human security, but also has a serious impact on economic growth and carbon emission.

What the post-epidemic carbon emission like? How to achieve established carbon emission target?

People have gotten through several crisis and accumulated some practical experience. When dealing with carbon emission change after COVID-19 pandemic, it may be helpful to take a lesson from 2008 global economic crisis. In addition, in order to promote carbon emission, it is necessary to monitor carbon emission changes and even go a step forward to figure out factors driving carbon emission changes.

Decomposition analysis is usually available to investigate carbon emission driving factors. Generally speaking, the structural decomposition analysis (SDA); the production-theoretical decomposition analysis (PDA); and the index decomposition analysis (IDA) are three frequently applied methods Zhang and Da, 2015) . These three decomposition methods all have advantages and disadvantages and applications (Chang et al., 2019; . In accordance with the purpose of this study, IDA is properly applied. Among various IDA methods, LMDI is J o u r n a l P r e -p r o o f 8 will worsen. Ahmed et al. (Ahmed et al., 2015) argued that the implementation of measures to cut down carbon emissions worldwide can only be achieved through the form of international trade. Trade openness helped more underdeveloped economies improved the national economic level and get rid of poverty. But the environmental pollution associated with this economic boom cannot be ignored (Ahmed and Long, 2013) . Consequently, more scholars invested in investigating the correlation between trade and environmental degradation. In essence, the contradiction between trade and environmental pressure is indirect, based on the fundamental theory that free trade stimulates economic growth and thus accelerates environmental degradation.

However, the impact of trade development on the environment has various performances in different countries, which may depend on the actual national conditions of each country, including economic development level, relevant economic policies, industrial structure and other aspects (Forslid and Okubo, 2015) . Given that, this chapter provides a brief review of existing literature on trade--carbon emissions.

From the perspective of single country, Andersson (Andersson, 2018) combined input-output frameworks and nonlinear models to explore the impact of trade liberalization on the rapid growth of CO2 emissions in China. Their results supported the fact that trade liberalization is the key determinant of the increment of carbon emissions embodied in import of China during the period 1995-2008. Based on the multi-regional output-output tables, Ren et al.(Ren et al., 2014) showed that the continued expansion of trade openness and the continuous inflow of foreign investment are important reasons for the soaring carbon emissions in China's industrial sector. It is worth mentioning that trade openness is seen as a pivotal indicator of inward direct investment, which is usually used to verify the authenticity of the PHH hypothesis. Actually, the PHH hypothesis depicts the long-term linkage J o u r n a l P r e -p r o o f 9 between carbon emissions, trade, and foreign direct investment. For instance, Farhani and Ozturk investigated the drivers of total carbon emissions in Tunisia from 1971 to 2012. The results were found to be against the PHH hypothesis, but the increase in trade openness would sacrifice environmental quality (Farhani and Ozturk, 2015) . Shahzad et al.(Shahbaz et al., 2017) found that Pakistan's trade openness and financial development led to an increase in carbon emissions during the period 1971-2011, illustrating that trade liberalization adversely affected the environment.

From the perspective of multi-region, some important economic organizations and emerging economies have aroused the interest of scholars, such as OECD, South Africa, BRICS and so on due to the increasing impact of unilateral trade on the free trade system. Managi et al.(Managi et al., 2009 ) investigated the impact of trade openness on carbon emissions in OECD countries and non-OECD countries and concluded that trade benefited the environmental improvement of OECD countries but accelerated GHG emissions in non-OECD countries. In order to verify the existence of the PHH hypothesis, Kearsley and Riddel(Kearsley and Riddel, 2010) focused on the impact of international trade on seven common environmental pollutants in 27 OECD countries. They concluded that a higher degree of trade openness was good for improving the environment and boost economic prosperity.

This study has made contributions to the relevant research of carbon emission changes in major two aspects. Firstly, in short-term research, this paper took 2008 global financial crisis as a lesson, tried to figure out possible carbon emission changes after COVID-19. Then, learning from previous experience, this paper exerted efforts to investigate factors driving carbon emission by applying LMDI decomposition analysis. Secondly, different from previous researches, this paper took trade into consideration and tried to uncover the impact of trade on carbon emission changes in a long-term research. On the whole, this paper examined carbon emission changes and influencing factors in both short-term and long-term, expecting to provide helpful references for carbon emission control after COVID-19.

According to Kaya identity, carbon emission (indicated by C) can be decomposed as follow: In accordance with LMDI decomposition model, the changes of carbon emission from base year to target year can be described as follows:

In Eq. (3), ∆ , ∆ , ∆ , ∆ demonstrate carbon intensity effect, energy intensity effect, affluence level effect and population scale effect. The calculation process is shown in Eq. (4) -Eq. (8).

In order to investigate the impact of trade activities on total carbon emissions and per capita carbon emissions, two indicators related to free trade: trade output and foreign direct investment are introduced. Additionally, per capita energy consumption, energy intensity and energy structure 1 are considered. Consequently, On the basis of the specifications of Shahzad et al. (Shahzad et al., 2017) and Zoundi (Zoundi, 2017), the long-term estimation equations of this paper are as follows: ln = 1 + 2 + 3 + 4 + 5 + 1 (9) lnper = 1 + 2 + 3 + 4 + 5 + 2 (10)

In the above formula, = 1,2,3, ⋯ represents the time span; represents the natural logarithmic form of the variables. represents carbon emissions at time t;

perC represents the per capita carbon emissions at time t; is the per capita energy consumption, is the energy structure, represented by the proportion of renewable energy in the primary energy consumption structure. denotes foreign direct investment, denotes trade output, calculated by the ratio of total import and export trade to GDP. is the long-time elastic coefficient between the influencing factors and the interpreted variable, and is the error term.

The stability of time series is the basic condition for conducting time series related research, and it is also a necessary prerequisite for ensuring the validity and reliability of empirical results. This paper uses the ADF unit root test method, which is widely used for examining the stability of raw time series, the regression model can be rearranged into the following form 

where = (∑ =1 ) − 1； = − ∑ = +1 ， c is a constant term. The Null hypothesis of the ADF test is 0 : = 0; that is, the time series contains a unit root and is unstable; the alternative hypothesis is 0 : < 0, indicating that the original sequence is stationary. Only when the null hypothesis is rejected can the time series be proved to be stable and suitable for modeling.

Johansen co-integration procedure is applied to detect the existence of a long-term co-integration relationship between variables. Compared with the traditional E-G co-integration approach, the Johansen test method has better J o u r n a l P r e -p r o o f 13 adaptability and accuracy. It can not only detect long-term relationships between multiple variables, but also obtain the number of co-integration relationships. The main principle of the Johansen test program is the loop test, which verifies whether the variables are integrated for a long time from the null hypothesis. The principle of Johansen co-integration program can be expressed in the following form:

In the above Eq. (12), ∆ represents the first-order difference form of variables, 0 represents the intercept term and the parameter of the equation; p is the number of lag periods. The null hypothesis of co-integration test ( 0 : 0 = 1 = 2 = 0 )

considers that there is no cointegration between study variables, and the alternative hypothesis ( 1 : 0 ≠ 1 ≠ 2 ≠ 0) indicates that the study variables are long-term integrated. It is only when the null hypothesis is rejected that the variables are long-term correlated.

The vector autoregressive (VAR) model is generally employed to detect or predict long-term and short-term dynamic correlations between economic variables (Xu and Lin, 2016) . It treats all variables as endogenous variables and overcomes the shortcomings of errors due to subjective settings in the simultaneous equation model (Cheng et al., 2019; Dolatabadi et al., 2018) . The general form of the VAR model can be expressed as follows:

Where Y denotes an × 1 endogenous vector, A represents the corresponding J o u r n a l P r e -p r o o f 14 × coefficient matrix. p is the number of lag periods of economic model, and is a random error term.

IR analysis is one of the most important analytical procedures in the VAR system.

It can comprehensively capture the impact of the impact variable on the response variable during the study period, reflecting the complex dynamic relationship between variables. The model can be rearranged as follows to perform an impulse response analysis: 

VD refers to decomposing the variance of an endogenous variable to other explanatory variables. VD can clearly demonstrate the contributions of each impact factor to the dependent variable, and then estimate their relative importance (Ahmad et al., 2017; Jadidzadeh and Serletis, 2017) . The dynamic VD technique proposed by

Sims (Sims, 1980) is defined as:

In the Eq. (16), each term represents the total effects of the j-th perturbation term on from the past to the present. Based on the above Eq. (16), this study J o u r n a l P r e -p r o o f assumes that does not have sequence correlation, then the variance of the variable can be expressed as:

As shown in the Eq. (17), this model uses the variance to estimate the total effects of the perturbation term j on the variable i from the past to the present. In this case, the variance of the variable can be decomposed into K different and unrelated effects.

As for short-term research, this study selects global economy and three income level groups (high-income, upper-middle income, lower-middle income) as objects to investigate the changes of carbon emission. The low-income countries are excluded because the energy consumption data is not available. Besides, data of carbon emission, energy consumption, GDP and population all come from the World Development Indicators released online by the World Bank(The World Bank, 2020).

In order to eliminate the impact of inflation, the GDP is constant in 2010 US$.

As for long-term research, this paper focuses on the multi-faceted effects of trade on carbon emissions. Five economic variables are used to explore the long-term effects J o u r n a l P r e -p r o o f 16 to GDP. To eliminate the interference caused by the heteroscedasticity of the raw data, this paper uses the natural logarithmic form of the annual data for calculation. All above data are collected from the World Bank(The World Bank, 2020).

As shown in Fig 1, Appliances to the Countryside" program, which aimed to stimulate consumption by providing subsidy. In this period, countries made economic recovery a top priority, while loosed environmental regulatory. Hence, energy intensity reversely deteriorated and caused carbon emission increase when recovering economy.

In addition, in late 2008, oil price was decreasing due to the declining demand.

As the crisis deepening, the oil price continued to go down. However, when economy tried to recover, more energy consumption will be needed. Low price oil will be a better choice out of economic reason. But as oil demand increase and economic recovery, oil price will also recover. in developing countries, using the labor and environmental resources of developing countries to produce, thus meeting the actual needs of the country .

In this process, along with the transfer of funds, developed countries have transferred part of the environmental pressure to developing countries. FDI is not conducive to developing countries to implement carbon emission reduction measures.

To sum up, EI has a positive impact on CE in MI group, while its impact in other two groups is negative. This result shows that EI of developing countries is still at a higher level, which may hinder their carbon emissions reduction. Compared with the other two groups, the countries in MI group should pay more attention to improving energy efficiency and accelerating the research and development and innovation of energy-saving technologies, thereby achieving the goal of reducing energy intensity.

ES negatively affected CE in all three groups. This indicates that the higher the proportion of renewable energy in the energy mix, the better the reduction of carbon emissions (Wang and Ye, 2017) . In view of this, strengthening the development and utilization of renewable energy, such as wind energy and solar energy, can accelerate the realization of carbon emission reduction targets.

J o u r n a l P r e -p r o o f 31 According to long-term estimates, TRD accelerates carbon emissions growth. Above results can be explained by the "Jevons Paradox" in energy economics (Yoo et al., 2019) . With the deepening of economic globalization, the international division of labor and the global production network become more complete, leading to closer trade links between countries. In this system, countries share talents, technologies and knowledge, expand trade openness, thus greatly reducing production costs and further promoting international trade. Under these circumstances, higher trade openness would promote carbon emissions rather than the reduction.

The model should be tested for stability before building, since the robustness of the model directly determines the accuracy and effectiveness of the experimental results. This paper uses the AR eigenvalue to detect the robustness of the model. It is verified that the AR roots of the variables are all within the unit circle, as shown in Figure 7 , indicating that the six economic models established are stable. trend. Finally, as the growth rate of EC slows down, its contribution to CE continues to decrease, and the image eventually stabilizes around zero (greater than zero). In a nutshell, EC is positively correlated with changes in CE because the release of GHG (Behera and Dash, 2017) . Consequently, reducing EC levels is helpful in controlling carbon emissions. From a long-term perspective, EI has a positive impact on CE, and this impact is more pronounced in the short term. Countries are trying to explore effective ways to cut energy intensity because of the widespread concern caused by climate change issues. Improving energy efficiency is considered effective in reducing energy consumption. It minimizes energy waste by improving and innovating energy technologies, thereby mitigating energy consumption in the production process.

ES negatively affected CE, indicating that the improvement in energy mix help cut down carbon emissions. In recent decades, countries are actively developing new energy and clean energy to replace traditional energy for production activities (Hansen et al., 2019; Liu, 2019) . As a result, these measures brought about an increasing share of renewable energy in the energy mix and slowed the accumulation of CE. FDI plays a negative role in promoting carbon emissions, although this hindrance is less significant. From a long-term perspective, FDI is conducive to improving the environmental quality of the country. Governments should encourage enterprises to introduce foreign capital and expand production scale, which can not only promote economic prosperity, but also prevent environmental degradation. TRD initially exerted a negative impact on CE, but since the third period, as trade share increased in GDP, it has a gradual weakening effect on CE. Although TRD has a positive impact on CE, this promotion is minimal and does not cause a significant increase in CE.

J o u r n a l P r e -p r o o f 34 Figure 9 -a. Impulse-response functions of CE to variables in HI Figure 9 -b. Impulse-response functions of perCE to variables in HI Figure 9 illustrates the IR images of HI. EC has a positive influence on CE throughout the study interval. It can be inferred that the EI of HI countries will cause carbon emissions growth in the future. Therefore, it is necessary for developed countries to take measures to showed an upward trend implying that EC will promote CE in MI countries in the future.

Similarly, the impact of EI on CE also shows obvious phase characteristics, with positive and negative effects alternating. These results indicate that EI in developing countries is still at a high level, which is a barrier to carbon reduction.

J o u r n a l P r e -p r o o f 36 The difference is that FDI can reduce carbon emissions in a short period of time, while TRD may promote CE in the short run. Therefore, although free trade is beneficial for developing countries to achieve CE reduction targets in the long run, and it causes rapid accumulation of CE in a short time. This finding supports trade liberalization.

In summary, in three groups, EC is positively correlated with CE. The impact of EI on CE varies among three groups. This mainly because HI countries have advanced energy technologies, which can ensure that the energy intensity is basically stable at a relatively low level and will not promote excessive growth of carbon emissions. But for MI economies, there is still much room for energy intensity reduction, so the promotion of carbon emissions is relatively significant. ES negatively affects CE, as renewable energy plays an increasingly important role in the energy mix. However, the impact of ES on CE is not significant, meaning that non-renewable energy still dominates. It is an effective way to solve this problem by vigorously promoting and popularizing clean energy. FDI has a negative impact on CE in the long run, although it causes CE to increase in the short term. This finding suggests that the establishment of the PHH hypothesis is conditional, that is, the large inflow of foreign investment in a short period will indeed damage the country's environment and bring about an increase in pollutant emissions; but in the long run, FDI will not lead to a sharp increase on CE.

The impact of TRD on CE is uncertain and depends on the country's income level. For developed countries, trade is conducive to carbon emission reduction, while for MI countries, trade negatively affects CE in the long run, but may accelerate it in the short term. There is not sufficient evidence in this paper that free trade is conducive to global carbon reduction.

This paper aims to explore carbon emission changes and factors influencing this change after this COVID-19 pandemic from both short-term research and long-term research. We have come to the following conclusions and proposed scientific and practical policy implications:

For short-term research, high-income level group initially achieved carbon reduction. As for middle-income level group, which significantly increased carbon emission, upper-middle income level group was deemed as the largest contributor.

Moreover, from the perspective of specific countries, it may come to a conclusion that developed countries were likely to curb carbon increase, while developing countries may still struggle with carbon emission control. In addition, since the impact of global economic crisis, most countries tended to slower or even reduce carbon emission in 2008-2009 and present a retaliatory rebound of carbon emission in 2009-2010, which may teach a lesson for carbon emission changes after COVID-19 pandemic.

Affluence level effect was prominent inhibitor to carbon reduction in all studies objects, particularly the upper-middle income level group and lower-middle income level group, which almost located in the process of rising economy. Energy intensity effect prominently drove carbon reduction, especially in high-income level group and G7 group. In these countries, the positive impact of energy intensity was stronger than passive impact of affluence level on carbon reduction. Consequently, improving energy intensity may also help to reduce carbon emission after COVID-19 pandemic.

More investment shall be put to promote energy-saving technologies and strengthen research and development in related technologies; more clean and renewable energy shall be used in present energy system; encourage more monetary and fiscal polies implicated to improve energy intensity.

impact on CE in middle-income level group, while negative impact on the other groups. It is essential to strengthen the research and development (R&D) and innovation of energy technology to reduce the energy intensity. Trade is shown to accelerate carbon emissions, but this promotion effect is minimal in the long-run. In comparison, international trade is more likely to contribute to the carbon emissions of middle-income countries.

According to the results of IR analysis, ES has a negative impact on CE, as renewable energy plays an increasingly important role in the energy mix, inhibiting the accelerated growth of CE. In all income groups, FDI is negatively correlated with CE in the long run but cause an increase in CE in the short term. This finding supports the PHH hypothesis under certain conditions. For developed countries, TRD is conducive to carbon emission reduction; for middle-income countries, TRD exerts a negative impact on CE in the long term but accelerates CE growth in the short term.

Consequently, government shall de devoted themselves to promote trade openness in long-term, since adhering to free trade is good for achieving global emissions reduction targets. 

Decomposition of Aggregate Energy and Gas Emission Intensities for Industry: A Refined Divisia Index Method

The effect of urbanization, energy consumption, and foreign direct investment on the carbon dioxide emission in the SSEA (South and Southeast Asian) region. Renewable and Sustainable Energy Reviews

Analysis: Coronavirus set to cause largest ever annual fall in CO2 emissions

Daily global CO2 emissions 'cut to 2006 levels' during height of coronavirus crisis

Driving forces of global carbon emissions: From time-and spatial-dynamic perspectives

The VEC-NAR model for short-term forecasting of oil prices

An LMDI decomposition analysis of carbon emissions in the Thai manufacturing sector

price fluctuations on Chinese stock market investor sentiment

Economic significance of commodity return forecasts from the fractionally cointegrated VAR model

Analysis: Coronavirus set to cause largest ever annual fall in CO2 emissions

Causal relationship between CO2 emissions, real GDP, energy consumption, financial development, trade openness, and urbanization in Tunisia

Which firms are left in the periphery? Spatial sorting of heterogeneous firms with scale economies in transportation

Full energy system transition towards 100% renewable energy in Germany in 2050

Global Energy Review 2020: the impacts of the Covid-19 crisis on global energy demand and CO 2 emissions

How does the U.S. natural gas market react to demand and supply shocks in the crude oil market?

A further inquiry into the Pollution Haven Hypothesis and the Environmental Kuznets Curve

Decomposition Analysis of Greenhouse Gas Emissions in Korea's Transportation Sector. Sustainability

Temporary reduction in daily global CO2 emissions during the COVID-19 forced confinement

Challenges for China's carbon emissions peaking in 2030: A decomposition and decoupling analysis

Low-carbon economic development in Central Asia based on LMDI decomposition and comparative decoupling analyses

Regional differences in carbon emission of China's industries and its decomposition effects

Carbon dioxide intensity and income level in the Chinese megacities' residential building sector: Decomposition and decoupling analyses

Driving forces of provincial-level CO2 emissions in China's power sector based on LMDI method

Consumption-based accounting and the trade-carbon emissions nexus

China's renewable energy law and policy: A critical review

Carbon emissions from energy consumption in China: Its measurement and driving factors

Does trade openness improve environmental quality

Analysis on the influencing factors of carbon emission in China's logistics industry based on LMDI method

International trade, FDI (foreign direct investment) and embodied CO2 emissions: A case study of Chinas industrial sectors

Paris Agreement climate proposals need a boost to keep warming well below 2 °C

A new scenario logic for the Paris Agreement long-term temperature goal

Trade openness-carbon emissions nexus: The importance of turning points of trade openness for country panels

Carbon emission, energy consumption, trade openness and financial development in Pakistan: A revisit

Macroeconomics and reality

How the coronavirus pandemic slashed carbon emissions -in five graphs

Embodied environmental damage in interregional trade: A MRIO-based assessment within

Is carbon emission growth decoupled from economic growth in emerging countries? New insights from labor and investment effects

Is energy transition promoting the decoupling economic growth from emission growth? Evidence from the 186 countries

Forecasting Chinese carbon emissions from fossil energy consumption using non-linear grey multivariable models

Provincial-level industrial CO2 emission drivers and emission reduction strategies in China: Combining two-layer LMDI method with spectral clustering

Study on The Influencing Factors of Agricultural Carbon Emission in Sichuan Based on LMDI Decomposition Technology

An analysis of the decoupling relationship between CO2 emission in power industry and GDP in China based on LMDI method

Reducing carbon dioxide emissions in China's manufacturing industry: a dynamic vector autoregression approach

Decomposing factors affecting CO2 emissions in Pakistan: insights from LMDI decomposition approach

Revisiting Jevons's paradox of energy rebound: Policy implications and empirical evidence in consumer-oriented financial incentives from the Japanese automobile market

Decomposition analysis of China's CO2 emissions (2000-2016) and scenario analysis of its carbon intensity targets in 2020 and 2030

Exploring the transformation and upgrading

The authors would like to thank the editor and these anonymous reviewers for their helpful and constructive comments that greatly contributed to improving the final version of the manuscript. This work is supported by National Natural Science